检索结果-内蒙古大学图书馆

On Open-Set, High-Fidelity and Identity-Specific Face Transformation

IEEE ACCESS 2020年 8卷 224643-224653页

作者： Zhang, Longhao Pan, Xipeng Yang, Huihua Li, Lingqiao Beijing Univ Posts & Telecommun Sch Artificial Intelligence Beijing 100876 Peoples R China Guilin Univ Elect Technol Sch Comp Sci & Informat Secur Guilin 541004 Peoples R China Guangdong Prov Peoples Hosp Guangdong Acad Med Sci Dept Radiol Guangzhou 510080 Peoples R China

In this paper, a Generative Adversarial Networks-based framework has been proposed for identity-specific face transformation with high fidelity in open domains. Specifically, for any face, the pro-posed framework can transform its identity to the target identity, while preserving attributes and details (e.g., pose, gender, age, facial expression, skin tone, illumination and background). To this end, an auto-encoder network is adopted to learn the transformation mapping, which encodes the source image into the latent representation, and reconstruct it with the target identity. In addition, the face parsing pyramid is introduced to help the decoder restore the attributes. Moreover, a novel perceptual constraint is applied to the transformed images to guarantee the correct change of the desired identity and to help retrieve the lost details during face identity transformation. Extensive experiments and comparisons to several open-source approaches demonstrate the efficacy of the proposed framework: it can achieve more realistic identity transformation while better preserving attributes and details.

关键词： Faces Generative adversarial networks Image reconstruction Training Generators Decoding Gallium nitride auto-encoder face transformation generative adversarial networks perceptual constraint

来源：评论

学校读者我要写书评

暂无评论

Adaptive Importance Channel Selection for Perceptual Image Compression

引用

KSII TRANSACTIONS ON INTERNET AND INFORMATION SYSTEMS 2020年第9期14卷 3823-3840页

作者： He, Yifan Li, Feng Bai, Huihui Zhao, Yao Beijing Jiaotong Univ Inst Informat Sci Beijing 100044 Peoples R China

Recently, auto-encoder has emerged as the most popular method in convolutional neural network (CNN) based image compression and has achieved impressive performance. In the traditional auto-encoder based image compression model, the encoder simply sends the features of last layer to the decoder, which cannot allocate bits over different spatial regions in an efficient way. Besides, these methods do not fully exploit the contextual information under different receptive fields for better reconstruction performance. In this paper, to solve these issues, a novel auto-encoder model is designed for image compression, which can effectively transmit the hierarchical features of the encoder to the decoder. Specifically, we first propose an adaptive bit-allocation strategy, which can adaptively select an importance channel. Then, we conduct the multiply operation on the generated importance mask and the features of the last layer in our proposed encoder to achieve efficient bit allocation. Moreover, we present an additional novel perceptual loss function for more accurate image details. Extensive experiments demonstrated that the proposed model can achieve significant superiority compared with JPEG and JPEG2000 both in both subjective and objective quality. Besides, our model shows better performance than the state-of-the-art convolutional neural network (CNN)-based image compression methods in terms of PSNR.

关键词： Image compression auto-encoder Perceptual loss Bit-allocate strategy Importance map

来源：评论

学校读者我要写书评

暂无评论

Deep-learning-based surrogate model for reservoir simulation with time-varying well controls

引用

JOURNAL OF PETROLEUM SCIENCE AND ENGINEERING 2020年 192卷 107273-107273页

作者： Jin, Zhaoyang Larry Liu, Yimin Durlofsky, Louis J. Stanford Univ Dept Energy Resources Engn Stanford CA 94305 USA

A new deep-learning-based reduced-order modeling (ROM) framework is proposed for application in subsurface flow simulation. The reduced-order model is based on an existing embed-to-control (E2C) framework and includes an auto-encoder, which projects the system to a low-dimensional subspace, and a linear transition model, which approximates the evolution of the system states in low dimension. In addition to the loss function for data mismatch considered in the original E2C framework, we introduce a physics-based loss function that penalizes predictions that are inconsistent with the governing flow equations. The loss function is also modified to emphasize accuracy in key well quantities of interest (e.g., fluid production rates). The E2C ROM is shown to be analogous to an existing ROM, POD-TPWL, which has been extensively developed for subsurface flow simulation. The new ROM is applied to oil-water flow in a heterogeneous reservoir, with flow driven by nine wells operating under time-varying control specifications. A total of 300 high-fidelity training simulations are performed in the offline stage, and the network training requires 10-12 minutes on a Tesla V100 GPU node. Online (runtime) computations achieve speedups of over a factor of 1000 relative to full-order simulations. Extensive test case results, with well controls varied over large ranges, are presented. Accurate ROM predictions are achieved for global saturation and pressure fields at particular times, and for injection and production well responses as a function of time. Error is shown to increase when 100 or 200 (rather than 300) training runs are used to construct the E2C ROM. Overall the E2C ROM is shown to provide reliable predictions with levels of perturbations in the well controls that are much larger than those used with existing POD-TPWL treatments. The current model is however limited to 2D systems, and the required number of training simulations is much larger than that for POD-based ROMs.

关键词： Reservoir simulation Reduced-order model Deep learning Physics-informed neural network auto-encoder Embed-to-control E2C

来源：评论

学校读者我要写书评

暂无评论

Multivariate Abnormal Detection for Industrial Control Systems Using 1D CNN and GRU

引用

IEEE ACCESS 2020年 8卷 88348-88359页

作者： Xie, Xin Wang, Bin Wan, Tiancheng Tang, Wenliang East China Jiaotong Univ Sch Informat Engn Nanchang 330013 Jiangxi Peoples R China

Currently, most anomaly detection approaches in industrial control systems (ICSs) use network event logs to build models, and current unsupervised machine learning methods rarely use spatiotemporal correlations and other dependencies between multiple variables (sensors/actuators) in a system to detect anomalies. Most of the existing anomaly detection technologies simply compare the current states with the predicted normal range. Due to the highly dynamic characteristic of industrial control systems, it is insufficient to simply compare the current states with the predicted normal range. As a result, these approaches have low detection rates for unknown or new types of attacks. In view of these shortcomings, this paper presents a network model for predicting sensor/controller parameters in industrial control systems. To predict the parameter values of the sensors and controllers more accurately, the 1D convolutional neural network (1D & x005F;CNN) and gated recurrent unit (GRU) are combined to fully learn the spatiotemporal correlation and other dependencies between the parameter values of the sensors and controllers at each moment. An abnormal state detection method based on the calculation of the statistical deviation is proposed to realize the anomaly detection of industrial control systems. The model is validated on the Secure Water Treatment (SWaT) dataset. The precision, recall and F1 scores are used to evaluate the effectiveness of this method in anomaly detection on the SWaT dataset. The experimental results show that the average precision and recall of this method are 0.99 and 0.85, respectively, and that the average F1 score is 0.91. The experimental results show that the proposed method can be successfully applied to anomaly detection systems in industrial control systems with lower false positive rates.

关键词： Industrial control Feature extraction Anomaly detection Intrusion detection Integrated circuit modeling Correlation auto-encoder 1D convolutional neural network gated recurrent unit industrial control system anomaly detection SWaT dataset

来源：评论

学校读者我要写书评

暂无评论

Attention-Based Event Characterization for Scarce Vehicular Sensing Data

IEEE OPEN JOURNAL OF VEHICULAR TECHNOLOGY

引用

IEEE OPEN JOURNAL OF VEHICULAR TECHNOLOGY 2020年 1卷 317-330页

作者： Taherifard, Nima Simsek, Murat Lascelles, Charles Kantarci, Burak Univ Ottawa Sch Elect Engn & Comp Sci Ottawa ON Canada Prod Dev Raven Connected 441 MacLaren St Ottawa ON K2P 2H3 Canada

Characterizing risky driving behavior is crucial in a connected vehicle environment, particularly to improve driving experience through enhanced safety features. Artificial intelligence-backed solutions are vital components of the modern transportation. However, such systems require significant volume of driving event data for an acceptable level of performance. To address the issue, this study proposes a novel framework for precise risky driving behavior detection that takes advantage of an attention-based neural network model. The proposed framework aims to recognize five driving events including harsh brake, aggressive acceleration, harsh left turn and harsh right turn alongside the normal driving behavior. Through numerical results, it is shown that the proposed model outperforms the state-of-the-art solutions by reaching an average accuracy of 0.96 and F1-score of 0.92 for all classes of driving events. Thus, it reduces the false positive instances compared to the previous models. Furthermore, through extensive experiments, structural details of the attention-based neural network is investigated to provide the most viable configuration for the analysis of the vehicular sensory data.

关键词： Attention model auto-encoder connected vehicles LSTM machine learning recurrent neural networks vehicular sensing

来源：评论

学校读者我要写书评

暂无评论

Multimodal Biometrics Recognition from Facial Video with Missing Modalities Using Deep Learning

引用

JOURNAL OF INFORMATION PROCESSING SYSTEMS 2020年第1期16卷 6-29页

作者： Maity, Sayan Abdel-Mottaleb, Mohamed Asfour, Shihab S. Univ Miami Dept Ind Engn Coral Gables FL 33124 USA Univ Miami Dept Elect & Comp Engn Coral Gables FL 33124 USA

Biometrics identification using multiple modalities has attracted the attention of many researchers as it produces more robust and trustworthy results than single modality biometrics. In this paper, we present a novel multimodal recognition system that trains a deep learning network to automatically learn features after extracting multiple biometric modalities from a single data source, i.e., facial video clips. Utilizing different modalities, i.e., left ear, left profile face, frontal face, right profile face, and right ear, present in the facial video clips, we train supervised denoising auto-encoders to automatically extract robust and non-redundant features. The automatically learned features are then used to train modality specific sparse classifiers to perform the multimodal recognition. Moreover, the proposed technique has proven robust when some of the above modalities were missing during the testing. The proposed system has three main components that are responsible for detection, which consists of modality specific detectors to automatically detect images of different modalities present in facial video clips;feature selection, which uses supervised denoising sparse auto-encoders network to capture discriminative representations that are robust to the illumination and pose variations;and classification, which consists of a set of modality specific sparse representation classifiers for unimodal recognition, followed by score level fusion of the recognition results of the available modalities. Experiments conducted on the constrained facial video dataset (WVU) and the unconstrained facial video dataset (HONDA/UCSD), resulted in a 99.17% and 97.14% Rank-1 recognition rates, respectively. The multimodal recognition accuracy demonstrates the superiority and robustness of the proposed approach irrespective of the illumination, nonplanar movement, and pose variations present in the video clips even in the situation of missing modalities.

关键词： auto-encoder Deep Learning Multimodal Biometrics Sparse Classification

来源：评论

学校读者我要写书评

暂无评论

Deep Neural Networks with Extreme Learning Machine for Seismic Data Compression

引用

ARABIAN JOURNAL FOR SCIENCE AND ENGINEERING 2020年第3期45卷 1367-1377页

作者： Nuha, Hilal H. Balghonaim, Adil Liu, Bo Mohandes, Mohamed Deriche, Mohamed Fekri, Faramarz King Fahd Univ Petr Ctr Energysignal ProcCeGPMineralsKFUPM Dhahran 31261 Saudi Arabia Georgia Inst Technol Sch Elect Comp Engn Atlanta GA USA

Advances on seismic survey techniques require a large number of geophones. This leads to an exponential growth in the size of data and prohibitive demands on storage and network communication resources. Therefore, it is desirable to compress the seismic data to the minimum possible, without losing important information. In this paper, a stacked auto-encoder extreme learning machine (AE-ELM) for seismic data compression is proposed. First, a deep asymmetric auto-encoder is constructed, in which nonlinear activation functions are used in the encoder hidden layers and linear activation functions are utilized in the decoder layers. Second, the encoder hidden layers are connected in a cascade way, so that outputs of a hidden layer are considered as the inputs to the succeeding hidden layer. Third, the optimal weights of connections between the layers of the decoder are solved analytically. Lastly, the AE-ELMs are stacked to create the complete encoder/decoder. The extreme learning machine (ELM) is selected due to its analytical calculation of weights efficient training that is suitable for practical implementation. In this neural network, data compression is achieved by transforming the original data through the encoder layers where the size of outputs from the last encoder hidden layer is smaller than the original data size. The proposed method exhibits a comparable reconstruction quality on a real dataset but with a much shorter training duration than other deep neural networks methods. This neural network with more than 8000 hidden units achieved 1.28x10-3 of normalized mean-squared error for 10:1 of compression ratio with only 8.23 s of training time.

关键词： Deep neural networks auto-encoder Extreme learning machine Seismic data compression

来源：评论

学校读者我要写书评

暂无评论

A GPU-based accelerated ELM and deep-ELM training algorithms for traditional and deep neural networks classifiers

引用

INTELLIGENT SYSTEMS WITH APPLICATIONS 2022年 15卷

作者： Chegni, Arezoo Moradi Ghavami, Behnam Eftekhari, Mahdi Shahid Bahonar Univ Kerman Dept Comp Engn Kerman Iran

The extreme learning machine (ELM) has been effectively used for training single-layer neural networks. In recent years, great attention has been paid to deep extreme learning machine (D-ELM) structures. Deep neural network structures are trained via the ELM method. Some stacked auto-encoders followed by a simple ELM layer can be usually used for solving classification and regression tasks. Although ELM has been employed for speeding up the training process of the neural network, D-ELM based models suffer from some issues such as the time complexity and running time. In this paper, we explore how the evaluation of ELM and D-ELM can be accelerated. GPUs are used to speed up the training process of ELM and D-ELM models. In the proposed method, three separate phases are considered for the algorithms. In the first phase, loading and pre-processing the data are performed serially in the CPU. In the second and third phases, which respectively are the training and testing phases of the algorithm, all the matrix operations of the algorithms are implemented in parallel mode using the GPU memory hierarchy. Also, having access to highly efficient computational libraries, additional support is provided for GPU-based parallel computing. In the simulation setup, five sets of the database are applied to train the ELM and D-ELM on both CPU and GPU platforms. The results obtained show the proposed approach based on GPUs can remarkable save running time. Although both serial and parallel methods measure approximately the same accuracy, the parallel methods provided for the models reduce the run time significantly.

关键词： Extreme learning machine Deep extreme learning machine auto-encoder CPU GPU

来源：评论

学校读者我要写书评

暂无评论

Conventional neural network for blind image blur correction using latent semantics

引用

SOFT COMPUTING 2020年第20期24卷 15223-15237页

作者： Gowthami, S. Harikumar, R. Dr NGP Inst Technol Dept Biomed Engn Coimbatore Tamil Nadu India Bannari Amman Inst Technol Dept Elect & Commun Engn Sathyamangalam India

In this work, deep learning for enhancing the sharpness of blurred image is investigated. Initial pre-processing is blur image kernel estimation which is critical for blind image de-blurring. In prior investigation, handcrafted blur features are optimized for certain uniform blur, which is unrealistic for blind de-convolution. To deal with this crisis, initially this work attempts to carry out kernel matrix estimation using latent semantic analysis (KME-LSA) in dermatology image. In order to enhance the image sparseness, this work modelled an image descriptor based on Gaussian mixture model in auto-encoder (GMM-AE) as a primary layer in convolutional neural networks. The functionality of the proposed GMM-AE triggers the selection of efficient features for subsequent layers in CNN. The features extracted from the integrated trained GMM-AE in CNN can fine-tune the quality of blurred image. Datasets used are melanoma-based dermascope images. Pre-processing procedures are carried out by LSA-based kernel matrix estimation. The attained sharp image outcome is given to the proposed model for effective feature extraction and to attain improved blind image. The anticipated KME-LSA and GMM-AE in CNN estimates blur parameters with high accuracy. Experiment illustrates the efficacy of proposed method and the competitive outcomes are compared with state-of-the-art datasets. Simulation was carried out in MATLAB environment;performance metrics like MSE-227.6, PSNR-33.6762, SSIM-0.9755 and VIF-0.08162 are evaluated. The results show better trade-off than the prevailing techniques.

关键词： Blur kernel estimation Blind image de-blurring Gaussian mixture model Convolutional neural network auto-encoder

来源：评论

学校读者我要写书评

暂无评论

Performance Evaluation of Deep autoencoder Network for Speech Emotion Recognition

引用

INTERNATIONAL JOURNAL OF ADVANCED COMPUTER SCIENCE AND APPLICATIONS 2020年第2期11卷 606-611页

作者： AndleebSiddiqui, Maria Hussain, Wajahat Ali, Syed Abbas Danish-ur-Rehman NED Univ Engn & Technol Comp Sci & Informat Technol Karachi Pakistan Karachi Shipyard & Engn Works Ltd Karachi Pakistan NED Univ Engn & Technol Comp & Informat Syst Engn Karachi Pakistan NED Univ Engn & Technol Elect Engn Karachi Pakistan

The learning methods with multiple levels of representation is called deep learning methods. The composition of simple but now linear modules results in deep-learning model. Deep-learning in near future will have many more success, because it requires very little engineering in hands and it can easily take ample amount of data for computation. In this paper the deep learning network is used to recognize speech emotions. The deep autoencoder is constructed to learn the speech emotions (Angry, Happy, Neutral, and Sad) of Normal and Autistic Children. Experimental results evident that the categorical classification accuracy of speech is 46.5% and 33.3% for Normal and Autistic children speech respectively. Whereas, auto encoder shows a very low classification accuracy of 26.1% for only happy emotion and no classification accuracy for Angry, Neutral and Sad emotions.

关键词： auto-encoder emotions DNN classification accuracy autism

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：